# Horopter beta-3 decisions, evidence and abstention

> Horopter beta-3 answers typed questions with evidence spans and abstention estimates. Its 0.4B and 1B weights have a non-commercial licence.

Updated 2026-11-04 (Australia/Adelaide)

By Fight Fire with Fire Robotics · Beta-3 release; first release 20 Oct 2026: 2026-11-04 · OPEN

An encoder-only decision model family built on Ettin, with 0.4B and 1B sizes. It returns typed answers alongside evidence spans, a relation to the state and an abstention estimate. It encodes and caches the state before seeing questions. Beta-3 adds non-decisive rule outcomes and a field-extraction adapter; the first release was 20 Oct 2026.

## Facts

- Size: 0.4B and 1B
- Base: Ettin encoders from jhu-clsp
- Licence: Weights, data recipe and docs: CC BY-NC-SA 4.0; code: MIT
- Inputs: English text or JSON state; up to 2,048 state tokens
- Access: Public weights; Python runtime; /v1/systemone, /v1/clairic and combined /v1/horopter endpoints
- Training restriction: Training includes TypeSafe [Jev](<https://decisionmodels.ai/models/jev>) API outputs. The card says TypeSafe terms prohibit using them to imitate its model or build a competing service.

## Claims and sources

- **VENDOR-REPORTED**: The maker reports ShARC four-way accuracy of 82.3 for the 1B beta-3 release, against 54.3 for its first release. ShARC training data is included, so the dev set is held out within the same distribution. On [JevBench](<https://decisionmodels.ai/jevbench>) hard, the 1B scores 40.5 against 64.0 for clef-flash. Generated-policy results do not establish the same performance on real rule text. [source](<https://huggingface.co/jmccardle/horopter-1B>)

## Links

- [Source: 1B model card and weights](<https://huggingface.co/jmccardle/horopter-1B>)
- [Source: 0.4B model card and weights](<https://huggingface.co/jmccardle/horopter-0.4B>)
- [Source: Runtime and recipes](<https://github.com/jmccardle/horopter>)

## Related projects

No related projects are listed in the registry.